Stable Emulation of an Entire Suite of Model Physics in a State-of-the-Art GCM using a Neural Network

TL;DR

This paper demonstrates a neural network emulator that replaces the entire atmospheric physics suite in a GCM, enabling faster and stable medium-range weather forecasts with maintained realism and stability over extended periods.

Contribution

The study introduces a single neural network emulator for the full suite of atmospheric physics in a GCM, achieving model speed-up and stability while preserving forecast accuracy.

Findings

Stable and realistic medium-range forecasts for 24 initial conditions.

Maintains stability in a year-long experiment and at higher resolution.

Preliminary results show promising accuracy and computational speedup.

Abstract

There has been a lot of recent interest in developing hybrid models that couple deterministic numerical model components to statistical model components derived using machine learning techniques. One approach that we follow in this pilot study is to replace an existing computationally expensive deterministic model component with its fast machine-learning-based emulator, leading to the model speed-up and/or improvement. We developed a shallow neural network-based emulator of a complete suite of atmospheric physics parameterizations in NCEP Global Forecast System (GFS) general circulation model (GCM). The suite emulated by a single NN includes radiative transfer, cloud macro- and micro-physics, shallow and deep convection, boundary layer processes, gravity wave drag, land model, etc. NCEP GFS with the neural network replacing the original suite of atmospheric parameterizations produces stable and realistic medium range weather forecasts for 24 initial conditions spanning all months of 2018. It also remains stable in a year-long AMIP-like experiment and in the run with a quadrupled horizontal resolution. We present preliminary results of parallel runs, evaluating the accuracy and speedup of the resulting hybrid GCM.